December 11, 2019

Anatomy of a Packet


The transportation of data across the Internet is achieved by breaking messages into packets. These are switched through routers until they reach their destinations. Once they have arrived at their destination, the packets are reassembled to form the original message. TCP/IP protocols implement the construction, addressing and reassembly of the packets.

A packet is a string of bits divided into different segments. At its core is a data segment or payload (part of the original message). Additional information is added in the form of headers or, occasionally, trailers. Packets pass up and down through a TCP/IP stack on their way to and from the Internet. Consequently, both the sender and the receiver must implement TCP/IP protocols. As the packet moves up the stack of the transmitter, each layer adds a header. Once it has reached its destination, the reverse process takes place. Each layer in the receiver will strip out the appropriate header in the received packet as the packet moves down the stack. Received packets are subsequently reassembled in the form of the original message.

IP headers contain information vital for delivery, including:

  1. The Internet address of the sending machine
  2. The Internet address of the destination machine

Packets persist for a short space of time before being discarded by a router or host. The header will specify this time; the maximum value is 255 seconds, and the current (2000) default value is 32. "Protocols specify things at a very detailed (and, to humans, tedious) level. But once they are agreed and tested, they can then become effectively invisible to us, the users, because computers handle them effortlessly in the background."


Next page » TCP/IP and packets

Previous page « ARPANET to Internet



















Up to top of page